image masking

I would like to be able to draw vector masks over images. (would be nice to also be able to copy and paste between images too)

so i can remove objects that are in some of the photos (moving boat) but not all.

secondly it would be nice to remove objects via the sparse cloud before meshing. could be a lot easier than masking 50 photos.

I had a go with this in fusion, but had issues losing exif data in the exported pngs. and i think it would be better done in rc, so you don’t lose all the extra info.

Agree. I actually asked this a couple of weeks ago.

Hi

The masking tool for ALIGNMENT will be ready in the next release (testing it if it works as intended)

For masking creation itself it is best to use an external application that has proper and efficient tools for this (Pshop, Nuke, etc.)

If you just need to solve a moving car, or so, it is not a big issue, as we “compensate” it to some degree in alignment. If you have sufficient overlap, then it is not an issue at all…

TBH it’s a pretty low priority for me compared to all the other stuff. Man, I can’t wait for the next release!!!

Actually program works Fantastic with moved parts on image and filter it. Also Im happy how it filter a statick backgroung and allign only by actual object. But using masks will be big profit.
I use mask in my Photoscan pipeline and will be cool to transfer it here.
For example, will be good to have Import mask as patern: image name + “_mask”
And creating mask from scan shape.

I see that the latest release says it has “Alpha masking in alignment”. Is that this kind of masking? How do we use it?

Hi Aaron Curtis

As for the use of Masking in alignment, you just need to prepare the mask in an external application and use TIF or PNG that can save masks and then you just load it in RC and it will use it. So you do not need to look for any settings in RC, if mask is present, then it will be used…

I think will be good to make some options how to use the masks and how to utilize it.
For example - right now I do not convert my RAW files to Tiff cause CR(CapturinReality) use this format very well. So I just save my time.
And mask can be done in a different way - color keying, difference, handmade(yes, some times its need to be done), and also backlights… so I need option to load masks from separated files.
Also different step sometime need different masks. For example than I shoot on the bright background I got bright bloom on model but I need to keep it to make geometry better. But then I start bake textures, I just mask boomed parts from image (I don`t want exclude image cause some information its still usefull and needed.)

Ah, ok, great! I think this is a good solution. For example, I can use batch processing in GIMP to alpha the sky in all images.

Using external soft to prepare mask for all images in set - its one more Step in pipeline. :frowning:
Please make ability to load masks files separately. I really need this.
Same way as Photoscan will be fine.

Hi RC Team,

are there any news on this topic.

I am fine to create the mask in After Effects or some other external program. However combining the source images with the alpha mask is another step in After Effects and adds the need of another format when using JPGs as source. I’d prefer to add the mask as separate file.

I assume this is what everyone who is used to this workflow is longing for.

I’d appreciate any update on this.

Thank you!

I also support this.

+1 maybe like competitor, seperate MaskFiles and Masks generated using 3DModel

So, 3 years has passed and no masking tool?

crazy that this feature after over 5 years is either not seen as important or not necassary. Please tell us why some highly requested features are not getting completed. Would like to know what the limitations are.

Image masks are available from release 1.0.3.10393

Tutorial: https://www.youtube.com/watch?v=xxsdbzBwWLE&ab_channel=CapturingReality&fbclid=IwAR1zS25CvInDdE77og1PTOM62NEQ2MdxD9Yhw3hLwNgy6OtIAiywhHth1_0

Do you have in plan to implement manual masking tool (like metashape, zephyr). Masks created from model is good, but not applicable in all scenarios.

 

e.g. I recently had this cultural heritage project (atrium), and function of building is museum. I could only take photos during working hours, so I had many things on my way (banners, props, …) which I had to move to had clean shots on walls and floor. There is no way that you could take photos with all the props out of the frame. So, after 2 days of processing more than 5000 photos, I have one wall with ghost effect of metal plate, that was leaning on the wall at some point.

 

This would be very easy to solve If we had build in manual masking. Even masking in PS would not be a problem, but as far I know, we can’t import mask to already aligned and calculated model (or I am wrong?!?!?).

Hi ArchCro, I added your vote to our feature request database. 

It is not possible to import masks into already computed model now.

Thanks,

I believe that this shouldn’t be much of a problem to implement.

On image tab you can add selection tools (laso, rect, circle, …) and option to add selection to mask and invert mask.

Option where to use mask already exist.